The quest for fresh sounds is eternal for a music producer of any genre. Just ask old-school legends such as Sir George Martin, Alan Parsons, Trevor Horn or Quincy Jones, who used highly inventive sounds in their work to captivate audiences. That is equally important today in our sonically competitive electronic- and urban-music landscape, where standing out on the radio or in a DJ set can make or break a track's success, and standing out often means having a really cool signature sound.
Years of dabbling in the art of editing samples to sound like retro gear has resulted in exciting new ear candy that's guaranteed to be fresh and unique in musical projects. That involves treating modern samples so they sound as though they're being fed through, or emanating from, classic circuits and electro mechanisms such as the Mellotron or early drum machines. However, it isn't about pure emulation; it's about paying respect to the vibe that certain vintage instruments produced when played. The goal is to infuse that quirky, characteristic sound quality and vintage charm into today's pristine (and some may say characterless) samplers.
For the most part, these tips are for multisamples, but they can also apply to one shots or beat loops. To explore them, you'll need to know your way around an advanced sampling application such as Native Instruments Kontakt, Apple Logic's EXS24, Steinberg HALion, Propellerhead Reason's NN-XT, Tascam GigaStudio or the like. Since you'll be performing some basic waveform manipulation, you'll also need a decent audio editor to hack away at individual files (the waveform editor in most DAWs should do the trick).
CAUGHT ON TAPE
Ostensibly the first sampler, the Mellotron emulated real-life sounds and instruments by playing back notes from prerecorded tape libraries of its now-famous strings, trumpet, “Strawberry Fields” flute, 8-voice choir, etc. Libraries consisted of 35 individual lengths of tape housed in a frame, with each piece corresponding to a note on the Mellotron's short keyboard. Every key had its own read head, so polyphonic playback was possible. Applying the sonic character of this rudimentary playback mechanism to modern sample libraries isn't difficult and can sound mighty special in a contemporary mix.
Begin by preparing the raw WAV files that make up the chosen multisample. Whether you've done your own multisampling or are using a third-party preset, create duplicate working copies of the original samples and drag them to a new folder because we'll be applying destructive edits to these files. For authentic-sounding playback, it's crucial that you work with a separate WAV file for each note. If your preset of choice uses pitch scaling between root samples, resample all scaled notes and save them as their own WAV files. Since the Mellotron sound is not velocity sensitive, only one sample per note is required; generally you would choose the highest velocity sample from velocity-switched stacks.
Next, to get the right tone going across the board, downsample each WAV to a 22kHz or 30kHz sampling rate (sometimes even lower) to achieve the typically limited 11kHz or 15kHz maximum frequency response of early tape. To retain smoothness and definition, don't perform any bit reduction. Then begin treating each sampled note separately on a case-by-case basis.
To mimic different timbres, playback qualities and artifact variances between the different tapes, I apply varying degrees of hiss (using noise plug-ins); playback head bump (high-peaking resonant EQ works great for that); momentary audio drop-outs (drawing volume curves into the waveform); intermittent crackles; and other oddball tape effects. Be inventive; pull out a few feet of an old blank cassette tape, crinkle it in your fist, respool it and record that section into your DAW or audio editor, picking the cooler glitches to paste/overdub into your individual multisamples. Subtlety is the key, so apply these effects sparingly and randomly to taste.
Now you trim the WAV files in preparation for mapping in your sampler. But first thing's first: The Mellotron's tapes were linear, not looped, so turn looping off in your sampler. There was no Velocity sensitivity or Aftertouch either, so nix those parameters as well. Second, the sounds on tape were only about eight seconds long, so keep all of your individual WAV files to that length, even shorter on some if you wish. If you hold a note past its length, it should abruptly stop, so you are forced to retrigger; that's the nostalgia of a Mellotron's static sound and is particularly cool in chords as notes drop off at slightly different times. I like to chop the attack transients at different places for each WAV to simulate the way each tape started slightly differently. For example, while most notes in an aaah choir patch might start with a nice, crisp aaa sound, try chopping others to begin more into the breathy ahh portion. It's usually a matter of only a few dozen milliseconds.
Finally, keymap each prepared sample chromatically across the keyboard in your sampler. It's cool to apply a gentle amount of slow-moving sine-wave LFO (0.05-0.1 Hz) and assign it to the sound's oscillator pitch to create a subtle pitch drift, which emulates the Mellotron's tape-feed variance across the keyboard. Another trick for tweaking different tape tones on each key is to EQ the mids and high mids differently for each key, with peaks of dirty-sounding resonance on some notes; you're looking for a variety of clear and muddy tones across the keyboard. Also, don't try to volume match from note to note; the odd loud or quiet note is required for that quirky Mellotron sound. Essentially, the goal is that no two keys should sound exactly alike.
Next time you need an electric keys sound, try something unique. Rather than reaching for your favorite Rhodes or Wurlitzer libraries, head straight for your acoustic grand piano samples. By applying a few basic tricks entirely within the sampler, you can “electrify” any piano sound, imparting the mechanical and tonal characteristics of an electric upon the harmonic and timbral familiarity of an acoustic.
The main aspects of an electric piano's tone come from the key mallet striking the tone fork and being amplified by the pickup. You can simulate mallet stiffness by applying a sharply curved 12dB/octave lowpass filter (to rid the acoustic's shrill high-end overtones) with cutoff frequency and resonance both controlled by Velocity, so that it really honks when you lean into it. The tone usually sounds like a mutation between a Wurlitzer 200A and a Yamaha CP-70 at that point.
The amp envelope of electric pianos is generally shorter than acoustics, so reduce the decay and sustain times to a level you prefer, ranging from that warm, open tone to a tight clunkiness. Add overdrive/distortion under heavy Velocity for good Wurlitzer-styled bite.
To achieve a more complex electric tone, try doubling the piano multisample and transposing the second part up a full octave. Brightly contour it with a velocity-controlled resonant highpass or keytracked bandpass filter, so that only a thin, ringy tone remains, and the frequency window being emphasized travels with your playing. Then, blending the mix between the two layers can give the effect of pickup symmetry. Adding tremolo, chorus/phaser/flanger, delay and other electric piano favorites brings it all to life.
Any modern sampled drum kit can be “retroized” to sound the part of early drum machines such as the E-mu SP-12/1200/Drumulator, Linn LM-1, Oberheim DMX or Akai MPC60. Each of these classics is a 12-bit sampler operating at a 40kHz sampling rate or lower, so you could start with the obvious and bit-reduce your samples. Downsampling a high-rate source sample, or recording a fresh sample at a low rate, is also crucial for obtaining that gritty, raw, lo-fi sound. The LM-1, for instance, was only 28 kHz, and the SP-12 was an even lower 22 kHz.
Because of the limited RAM sizes and sampling time in each of these machines, owners would often speed up their sound sources (records, tape, etc.) and sample them in at a faster speed to consume less memory, only to pitch the samples down to normal speed inside the drum machine. This natural compression contributed to the punchy sound associated with these classics. This process is also what gave boxes like the DMX a heavily grained, almost broken-up character in the low end. You can achieve the same effect by pitch-bending (not time compressing) a sample up, recording it as a new file, importing it into your sampler and coarse tuning to its ideal pitch once it has been keymapped. (Tip: It's better to do that before bit reduction and downsampling.)
Sound designers would also severely truncate their samples so that an entire kit's worth of sounds could fit onboard. Some producers even liked to sample drums preprocessed with a touch of reverb before truncating them void of any reverb tail. The byproduct of that was a hollow, boxy sound that became popular for early rap recordings. Also try manually phasing/flanging things like cymbals, hats, claps and so on using two of the same samples atop one another — offset by just a few milliseconds — possibly micro-pitching them out.
The ARP Solina String Ensemble is often considered the definitive sound of the late '70s disco era. Although its trademark violin, viola, cello and contrabass sounds were based on analog sawtooth oscillators, I like starting with crisp, clean and harmonically rich modern symphonic string multisamples and degrading them for a rougher, edgier tone, which is awesome for trippy lo-fi hip-hop hooks. To create an ensemble, I'll typically work with two layers in a sampler: one for a solo violin or small string section and another containing the same multisample transposed up an octave (for those classic high, hanging disco lines) or a cello/contrabass pitched down an octave.
The secret to the Solina sound is to start with a fairly thin and nasal tone, as though coming from an old gramophone recording; otherwise, the patented heavy chorusing to be applied later will have no room to actually fatten up the sound. Scout around for a smooth, sustained sound with lots of rosin on the bow for a nice, authentic bite that you can “synthesize down” using filters. Immediately collapse any stereo samples to mono, reduce the bit-depth of each source wave to as low as 8 or 12 bits depending on how noisy and grainy you want it, reset loop points as necessary and apply highpass filtering with a gentle 12dB slope cutoff around 800 Hz. To retain detail and high-frequency options later, I prefer to leave sampling rates untouched in this case, instead using real-time lowpass filter controls in the sampler to contour the high-frequency curve.
The final step is to modulate each part against one another so that they phase into a remarkably thick and moving ensemble. Whether you've built a two-, three- or four-part sound, use slightly offset LFOs to control the pitch (which sets the phase) of each layer's sample-playback oscillator, and try assigning those same LFOs to also modulate the cutoff and resonance setting of a notch filter on each layer for extremely lush, sweeping pads.
Bit-depth — sets a sampler's dynamic range as well as gradation thereof (number of amplitude steps). Lower bit depths produce noisier, grainier and less defined sound, especially during softer passages, which can sound particularly cool on strings, for example.
Decimation — often erroneously referred to as downsampling, it is the process of applying lowpass filtering to a signal followed by throwing away some of its samples.
Downsampling — only throws away samples, skipping the lowpass-filtering operation of decimation.
Gate/truncate — gating is a process that acts upon volume dynamics and was popular in creating all those gated reverb drum presets in the '80s; truncating is a process of time and was often used to force-fit samples of precise length into early drum machines with very limited RAM.
Mono/stereo — collapsing stereo source samples down to mono for authentic vintage monaural sound can be advantageous if retention of program material from both sides is important, but sides can also nullify if frequencies cancel out, so be careful.
Sampling rate — defines the sampler's bandwidth or frequency response; in accordance to Nyquist's theorem, sampling rate is typically perceived as the “brightness” quotient of a sampled sound. The higher the sampling rate (and bit depth), the more accurately the original sound can be represented.